elections after the coup d'etat.pdf

7/27/2019 Elections After the Coup d'Etat.pdf

1/20

Elections After the Coup dEtat:

The International Community and the Seizure of Executive

Power

Hein Goemans

[email protected]

University of Rochester

Nikolay Marinov

[email protected]

Yale University

January 21, 2011

Abstract

In this pap er, we use new data on coup detats and elections to uncover a striking change in what

happens after the coup. Whereas the vast majority of successful coups before 1990 installed their

leaders durably in power, between 1991 and 2001 the picture reverses, with the majority of coups

leading to competitive elections in 5 years or less. We argue that with the end of the Cold War,

outside pressure has produced a development we characterize as the electoral norm - a requirement

that binds successful coup-entrepreneurs to hold reasonably prompt and competitive elections upon

gaining power. Consistent with our explanation, we find that post-Cold War those countries that

are most dependent on Western aid have been the first the embrace competitive elections after the

coup. Our theory is also able to account for the pronounced decline in the non-constitutional seizure

of executive power since the early 1990s. While the coup detat has been and still is the single most

important factor leading to the downfall of democratic government, our findings indicate that the

new generation of coups have been considerably less nefarious for democracy than their historical

predecessors.

Second draft. First draft presented at ISA 2008 and Midwest 2008. Preliminary and incomplete. We thank Jay Ulfelder

of the Political Instability Task Force for earlier comments. Any remaining errors are ours alone.

1


2/20

1 Introduction

I came in on a tank, and only a tank will evict me. - Abu Zuhair Tahir Yahya, Iraqi Prime

Minister, 19681

The first measure will be to to recall the previous parliament and make sure the proceedings areconstitutional. - Muhammad Naguib, leader of the Egyptian Revolution of 19522

The military coups in Greece in 1967 and Mali in 1991 began alike in some ways. The ruling au-

thorities had become deeply unpopular. The plummeting appeal of the government ushered political

instability. In Greece, the crisis took the form of a prolonged stand-off between the king and ousted

prime-minister George Papandreou. After months of domestic political turmoil, culminating in street

protests, the army moved to restore political order. In Mali, the long-time one-party government of

Moussa Traore refused demands for reform and fired at protesters. In both cases, the army succeeded

in gaining power in a quick putsch. But the two cases ended differently. In Mali, the army decided to

give up power. The country held competitive elections one year after the coup. The Greek colonels,

in contrast, chose to stay in power. It took a botched foreign intervention, and the specter of war

with neighboring NATO ally Turkey, to dislodge the military from office.

The time it takes a country to return to elected government after a coup has largely been neglected

in the literature. Scholars have focused their attention on the origins of the coup detat and the

consequences of military rule. Among the questions asked have been what types of countries are

at risk of experiencing coups,3 whether coup-plotters seek to promote a particular ideological or

corporatist agenda once in power, 4 whether coups cause lower economic growth or are caused by

economic hard times.5

In this paper, we uncover substantial variation in the time it takes a country to return to civilian

administration after a successful coup. Notably, the variation breaks down by historical period.

Before the end of the Cold War, almost all forceful seizures of power resulted in the military keepingpower to themselves. After 1990, the ma jority of coups committed lead to competitive elections.

Paradoxically, especially when compared to their predecessors, contemporary coups are a step toward

rather that away from democracy.

What caused this remarkable reversal? Why would a political actor powerful enough to seize

power by force be willing to give it up?

We argue that external factors drive significantly the decision to hold elections after a coup.

With the end of the Cold War, the West has began to promote free elections in the rest of the world.

While elections have not always been free and fair, nine of every ten countries in the world today

hold regular elections that are significantly more competitive than the forms of political contestation

most of these countries had before 1990.6 Furthermore, Western pressure to hold elections has been

felt especially strongly by the set of countries most likely to undergo a coup: countries with weak,

underdeveloped economies, and with po orly developed or unstable domestic authority structures.

To help us study the calculus of coup-plotters we present a simple two-stage optimization problem

with two related but distinct choices: whether to attempt to seize power and what to do with power

if successful at seizing it. The two-stage representation is helpful because we get to observe what

happens after a coup only if one takes place, and the factors that may influence the choices actors make

1Quoted in Luttwak (1969, p. 149).2Quoted in Finer (1988, p. 32).3See, for example, Jackman (1978), OKane (1981), Johnson, Slater and McGowan (1984), Wang (1998), Belkin and

Schofer (2003).4Jackman (1976), Zuk and Thompson (1982).5Londregan and Poole (1990), Alesina et al. (1996).6Authors data.

2


3/20

at each stage may be related. In our setup, coup plotters care about the attitude of the international

community. Less developed countries are especially susceptible to coups but they are also most likely

to listen to outside donors if asked to hold elections. When the international community imposes a

requirement to hold elections on those committing coups, it is especially likely to be heard.

Our empirical section uses data on foreign aid as a proxy for Western pressure to hold elections

post-1990. We find results consistent with our argument. We show that dependence on Western aid

tends to make countries more likely to hold competitive elections after coups - but this result only

holds for the post-Cold War set of cases. We get no relationship between aid and the speed with

which elections are adopted for the period between 1960 and 1990.

Our argument seeks to add to our understanding of democracy and the role of the outside world

in promoting democratization in three ways.

First, while it is conventional in the literature to equate the coup detat with lapse in democratic

institutions,7 we show that this is more of a special case than a general empirical regularity. When the

international community has ambiguous or no commitments to democracy, coups occur in weak states,

where it is it easy for coup-plotters to grab power and then keep demands for re-democratization

through elections at bay. Hence, the paucity of election events after coups in the pre-1990 period isnot surprising. A change of international attitudes in which democracy moves up the agenda alters

the picture significantly. Coups may be associated with authoritarian transitions - but they may also

be associated with pro-democratic outcomes.

Second, the dramatic expansion of democratic countries in the world in the course of the third

wave of democratization has been accompanied by a steady drop in the average per capita income

of the countries constituting the democratic club.8 Our work on coups suggests why representative

institutions may have survived better than expected in conditions where they might otherwise be

expected to quickly falter. Historically, the coup detat has been the biggest single danger to democ-

racy, accounting for 3 out of every 4 lapses in democracy according to one authoritative source.9

By discouraging coup-plotters, international forces have facilitated the stability of democracy in pre-

cisely those cases where representative institutions need help the most - in weak and underdeveloped

states. This finding contributes to the vibrant literature on the role of international factors for the

emergence and survival of democracy (Pevehouse, 2002; Dunning, 2004).

Third, the emergence of the electoral norm helps explain (among other factors) the over-time

drop in coup activity we uncover in the data. An important implication of our two-stage model is

that we expect the externally-imposed requirement of holding competitive elections after a coup to

reduce the incidence of such events in the first place. The electoral norm reduces the incidence of

coups by making the prize of gaining power less attractive to coup-enterpreneurs. In 2004, there

were 21 countries headed by leaders who came to power through a coup - a significant drop from the

historical maximum of 42 such states in 1976. Notably, this drop cannot be attributed to advances

in economic development as the best known theory of democratization - the modernization thesis- may lead us to expect (Lipset, 1959). The non-Western world, with few exceptions, is about as

economically developed today as it has been at the height of the Cold War. Thus, the decrease in

coup activity has come while the number of poor states has stayed relatively steady.10

7It is telling that in formal models of democratization and dictatorship, the coup represents the event of taking powerfrom the median voter and transferring it to the propertied elite (Acemoglu and Robinson, 2006).

8The pre-1990 democratic club had an average income of 11,510 dollars as compared to 8,883 for members of the sameclub after 1990 (constant 1995 dollars).

9The authors map their (annual) coup data against the democracy-autocracy dataset by Cheibub and Gandhi (2004)to derive this number.

10The average income of countries outside the West in 1975 was 2,474 dollars, as compared to 2,358 average income inthe same group in 1995 (figures in constant 1995 US dollars).

3


4/20

2 Theory

While there has been a lively literature in political science on the causes of the coup detat, it

provides few leads on whether or not, and with what speed, competitive elections may fallow the

grab of executive power.

Analytically, the problem of the coup detat is usually viewed as a problem of political instability.Political instability, whether manifested in institutional gridlock or mass protests, invites members

of the ruling elite or military to supplant the government and take the reins of power in their own

hands. There are many explanations of the causes of political instability but two stand out. One

stream of thought examines a countrys rate of economic modernization and its rate of political

development, and traces the roots of instability to lack of congruity between the two. Karl Deutsch

famously observed that as economic modernization transfigured urban and rural communities around

the world, it created immense pressure on governments to meet the demands of a new class of

politically conscious individuals.11 Sometimes governments would fail to reform fast enough to meet

the challenge, and political instability would follow. In a similar vein, Samuel Huntington noted

that economic modernization could ultimately transform traditional societies into stable polities but

the process itself may be profoundly destabilizing. Political instability, according to Huntington is

highest for societies that have embarked on a process of economic modernization but have not yet

developed the political institutions of economically advanced polities.12

A related but distinct set of arguments views the problem of political instability as a problem of

political legitimacy. Governments become illegitimate as they fail to deliver on the expectations of

the citizens. Economic performance is an important measure of how governments are able to meet

expectations. Thus, economic decline can be profoundly destabilizing while economic growth may

solidify a governments claim to legitimacy.13

Other explanations of the coup detat focus on how and why the army - the most likely culprit

in the event of one - intervenes in politics. Often the army sees itself as the bulwark against chaos,

and justifies its intervention in government with reference to real or alleged threat to a countrysinstitutional stability, economic welfare, or foreign policy direction.14 Whether the army will be

successful or not in seizing power is thought to depend on many factors. It has been suggested that

success may be more likely where the army holds a special place in society, due, for example, to

events surrounding the origin of the state. The army and its leaders played a central role in the

founding of Turkey, and in the liberation wars that spawned states in Latin America.15

Existing theories have helped us made headway both empirically and theoretically. For example,

one of the robust findings in the literature is the link between economic backwardness, economic

performance and coup activity - societies at a higher level of economic development or countries with

robust economic growth are systematically less likely to experience a forceful seizure of executive

power.16 This is consistent with versions of the modernization and legitimacy arguments.

Whether we believe that incomplete modernization or poor p erformance is at the root of the coup

detat, it is not clear what that should lead us to expect about the likelihood of elections after a

successful grab of executive power.

This is an important omission from an empirical point of view. We know that debates about

what to do with power are common among successful coup-leaders, and that they result in different

11Deutsch (1961).12Huntington (1968).13This argument is widely made, for versions, see McGowan and Johnson (2003) and Przeworski and Limongi (1997).14Ethnic diversity is often cited as a threat to a countrys unity, triggering coups (Johnson, Slater and McGowan, 1984).

The breadth of civilian control over the military is cited as the main explanation for why coups occur by Trinkunas (2005).15See Cheibub (2006) on the link between historical legacy and coup-proneness.16See Londregan and Poole (1990), Johnson, Slater and McGowan (1984).

4


5/20

outcomes. In the 1960 Turkish case, the army moved to take away power from the inept and

increasingly authoritarian Menderes government. Soon, disagreement emerged between the armys

commander in chief, general Gursel and some the younger brass. The General favored a quick return

to civilian administration while others wanted the army to stay in power longer to reform Turkey. In

the end, the pro-election faction prevailed.17 After the 1962 coup detat in Peru, ranking officer of

the military junta, General Ricardo Perez Godoy favored a return to negotiations with the elected

Congress. His viewpoint lost to younger members of the junta who wanted to remake the political

composition of the elected legislature (Needler, 1966). A similar rift within the Egyptian Free Officer

movement emerged between the formal leader of the army General Muhammad Naguib and the young

charismatic Nasser soon after the coup on July 23, 1952.18 Ultimately, Nasser won and elections lost.

Theories of democratization provide some traction on the question of post-coup elections. A

well-known proposition in the literature states that democratization and economic development are

systematically related (Lipset, 1959). While the exact nature of the causation is subject to debate,19

one plausible hypothesis to arise from that research would link the holding of competitive elections to

a countrys level of economic development. A more developed country is more likely to face pressure

for elections, all else equal.A different way to approach the question of post-coup elections is to ask about the role of the

international community in swaying the choices successful coup-leaders make. With some exceptions,

the role of international factors has been neglected.20 This is unfortunate. We know that coup-

leaders care intensely about international reactions. In his first communique, Colonel Bokassa of

the Central African Republic hurried to announce, among policy changes such the abolishment

of the bourgeoisie, one important continuity - that all foreign agreements shall be respected.21

The Greek colonels, having seized power in an almost bloodless coup in 1967, intensely lobbied the

U.S. government for speedy recognition.22 General Manuel Orellana, who ousted the Conservative

government in Guatemala in 1931, tendered his resignation before the National Congress after the

U.S. withheld recognition. International support or opposition to coups is both consequential and

variable.

How quickly, if at all, would successful coup leaders move to hold competitive elections? Is the

decision to grab power related to the choice of what to do with it after the event? What is the role

of international forces and factors at the two stages?

To help answer these questions, we consider the following simple time-line of a possible coup-event.

1. A coup plotter chooses between attempting a coup or sticking with the status quo. If the status

quo is chosen, the actor gets tsq [0, 1] in expected utility, which can be thought of as the

benefit of some policy outcome or as a transfer of resources.

2. If attempted, the coup succeeds with probability (0, 1). Failure yields 0 in expected utility.

3. If successful, the coup enterpreneur decides between calling for elections or retaining power.Calling for elections brings expected utility of tm [0, 1], which is simply where the median

voter would set the policy outcome or the transfer of resources.

17See Finer (1988, p.33) and Yalman (1968).18See LaCouture (1970, p.100).19See, for example, Przeworski and Limongi (1997) and Boix and Stokes (2003) on the question of whether income

facilitates democratization or merely enables democracy to survive.20Some of the scholars who discuss the international dimension are Coufoudakis (1977), who talks about the impact

of the European Economic Communitys decision to suspend the Greek association agreement on the juntas behavior;Hagopian and Mainwaring (2005), who illustrate the importance of changing North American attitudes under Carter forLatin America; Le Vine (2004), who discusses the French role in supporting and thwarting coups in Francophone Africa.

21See Luttwak (1969), p.175.22State Department, Memorandum regarding U.S. policy toward Greece following the 4/21/67 military coup in that

country (issued: Feb 27, 1968; declassified: Mar 15, 1996).

5


6/20

4. If the coup-plotter attempts to stay in power, she or he succeeds with probability (0, 1).

Failure yields 0 as a payoff; success yields a payoff of 1.

To generate comparative statics, we assume that and are themselves functions of a parameter

tapping dependence on foreign-provided b enefits i, and of a vector of other parameters w. We will

assume that and are continuous and second-order differentiable functions ofi and, for simplicity,

that the second-order derivative of i is 0.

The actor will attempt to keep power after a coup if:

tm 0

A coup will be attempted if:

V tsq 0,

where V, the expected continuation payoff after a successful coup, is:

V = if t

m 0,tm if tm < 0.

The main feature of this model relative to what has been done before is that it explicitly endows

the coup-leaders with a choice to hold elections after a successful coup. Incentives to move to

representative government exist regardless of the presence of international pressure on the country.

Above, elections will be attractive if the probability of staying in power through repressive means ()

is low relative to the policy outcome the coup-actor will get if they put the median voter in charge

(tm). Second, the attractiveness of a coup increases if the coup-plotter would benefit from moving

policy from the status quo to the median voters preferred point (tm > tsq). This implies that the

more out of tune the status quo is with what the median voter would favor, the more likely a coup

may be.This simple model is flexible enough to allow us to think about the role of different domestic

factors in the decision to take power by force. Take, for example, what has sometimes been called

the guardian coup: the military takes power from a corrupt and inept civilian administration, and

promises to return the country to elections after reforming the system to ensure truly representative

government. The 1960 coup in Turkey, the recent Bangladeshi coup in January of 2007 fit this

description. In such cases, the actions of the army are greeted by enthusiasm, as is the promise to

hold fresh elections quickly after purging corrupt politicians. When unpopular governments move

status quo policy away from the median voters ideal point (through corruption or repression against

the opposition), coup-leaders may have enough to gain from taking and then relinquishing power in

terms of policy (tm tsq) to make the grab of power and subsequent elections worthwhile.

The main comparative static of interest is the impact of international pressure on the decision to

perpetrate a coup and hold elections after.

Turning to the role of international pressure, for a value of the argument i = i and for a specific

draw of the parameters w = w, the marginal effect of an increase ofi on the attractiveness of keeping

power after a coup is:

V(i, w)

i

tmi

=V(i, w)

i(1)

The marginal effect of i on whether to attempt a coup depends on the arguments impact on the

probability of successfully grabbing and retaining power:

6


7/20

i[V tsq] =

(w, i)

iV(w, i) +

V(w, i)

i(w, i) (2)

Equation (2) helps to make clear that the international community has two distinct ways in which

it can alter the calculus of coup-plotters: by making it potentially harder or easier to seize power fromthe government in office ((w, i)/i), or by insisting on elections after a coup (V(w, i)/i < 0).

The sign of the respective derivatives may be positive, negative, or 0. The sign would probably

differ according to the specific circumstances of the country in question, its policies, and the relations

with the donors. Still, the case can be made that different average effects can be associated with

broad historical periods.

During the Cold War, the United States and the former colonial powers in Europe had an am-

biguous attitude toward coup plots: sometimes helping, sometimes thwarting, and sometimes doing

nothing. The American involvement in Allendes removal has given U.S. democratization policy of

the period a po or reputation. In reality, the United States sometimes opposed coups in support of

freely elected governments.23

After the Cold War, major players in international affairs, including the United States and the

European Union, have professed a commitment to defend democracy, including by punishing attempts

to bring down elected incumbents. In fact, since 1997, the US President has been bound under an

act of US Congress to suspend foreign aid to another country in the case of a coup detat.24 A

comparable commitment has been made on the EU level in 1991.

Western concern with coups has been part of a broader pattern of promoting competitive elections

in the developing world. After 1989, the United States, former colonial powers Britain and France,

the European Community, among others, converged on the idea that power around the world needs

to be shared through the ballot box. While this commitment has not been consistently enforced

everywhere, and the result has not always been liberal democracy, democratic conditionality has

helped make elections a nearly universal political ritual.Arguably, the conditioning of outside resources on elections has important implications for ac-

tors contemplating whether to seize power. While the events surrounding the execution of a coup

detat occur often too quickly for direct Western intervention to make a difference, the aftermath

of a successful seizure of power presents the new leaders with a difficult and potentially protracted

consolidation phase. The international community is presented with ample opportunities to press

conditions on the countrys leaders, while they worry about having sufficient resources to stave off

challenges to their untested grip on power.

Not surprisingly, recent coup leaders have often made negotiations with Westen donors a top

priority upon seizing office. The first announcement often is a commitment to institute or restore

consitutional order and elections. Major Daouda Malam Wanke, head of the presidential guard that

deposed the corrupt leader of Niger in 1999, assured the European Union that elections will be held

soon and requested a life-line of Western aid. One of the p oorest countries in the word, Niger lives

off foreign aid - some 80 % of its operating budget. The September 2003 coup in Guinea-Bissau, the

coup in Mali in 1991, among others, tell similar stories.

These observations suggest that the international community had an ambiguous commitment to

preventing coups or to instituting elected government to succeed juntas before 1990. Formally, this

may be represented by assuming that, on average, (w, i)/i = (w, i)/i = 0. This is the

case where international involvement may be present but can go in any possible direction.

23President Kennedy, for example, opposed the army takeover in Peru in 1961 (Naftali, 2001, p. 37).24See 508, Foreign Operations, Export Financing, and Related Programs Appropriations Act, 1997 (101(c) of title I of

Public Law 104-208).

7


8/20

As for the post-Cold War world, we would expect the effect of i on keeping power through

repressive means to be negative, /i < 0. This would make elections more attractive after a coup

for those governments that actually depend on foreign benefits.

The effect of international pressure on , the probability that a coup detat will succeed, is

somewhat more ambiguous. While we may want to say that Western opposition to coups translates

into a negative effect on , one may also argue that the leverage outsiders have in influencing the

rapid chain of events associated with a power grab is limited. It is plausible to assume for post-1990

either (w, i)/i < 0 or (w, i)/i = 0.

If the effect of the end of the Cold War on and can be summarized in this way, then revisiting

conditions (1) and (2) yields predictions on coups and post-coup elections. International pressure

makes it more likely that we will observe elections after a successful coup but it also tends to make the

prize of gaining power less attractive. In that way, it also so reduces the likelihood of coups. At some

level, this creates a sample attrition problem in observable data. In some cases where international

pressure would have produced elections, we would not observe a coup. At another level, this type of

sample selection would bias an empirical test of foreign dependence on post-coup elections against

finding the hypothesized effect.Conditioning of outside resources on elections will make some coup-leaders who would have other-

wise assumed dictatorial power switch their choice to holding competitive elections. If the distribution

of the parameters is such that those types of leaders continue to be interested in committing a coup

part of the time, then we should be able to uncover a relationship between dependence on outside

resources and the holding of post-coup elections. If there is sufficient residual variation among the

countries that do experience coups in terms of foreign dependence, we should be able to detect the

role of outside pressure on the propensity to adopt competitive elections.

Among the other implications of the hypothesized changes in the derivates is that we should see

fewer coup detats, both because success at gaining power may be somewhat harder to come by and

because having to hold elections will yield less in expected benefits.

Finally, we should see the number of elections after coups increase. Changing international

attitudes may encourage some actors to hold elections, and will not reduce the incentive of any

successful coup-plotter to place policy with the median voter.

H1 Competitive elections are more likely to follow a coup detat after the end of the Cold War

H2 Greater dependence on Western benefits should make post-coup elections more likely after the

end of the Cold War

H3 There should be a decline in the incidence of coups after the end of the Cold War

This paper focuses primarily on testing Hypothesis 2, but will provide some empirical support

for Hypotheses 1 and 3.

The simple model presented above does not incorporate all factors that may bear on the decision

to seize power by force. Yet, it is flexible enough to allow us to think about other factors at play.

For example, any factor that increases both the prospects of successfully deposing the government

and retaining power afterwards can be substituted for or added to the argument for international

pressure i currently in the model. Ethnic heterogeneity, past experience with coups, colonial history

are specific examples.

3 Data and Definitions

We want to know how long it takes a country to adopt elections after a successful coup detat.

Therefore, we need data on coups and on elections.

8


9/20

We begin by defining what we mean by a coup detat, then present our data on coups. After

that, we define the set of elections we are interested in, and present our data on elections. The we

outline our identification strategy.

The Coup DEtat

The coup detat is defined here as the seizure of effective executive authority through the threat or

use of force. The actors perpetrating the coup may include the military, the police, a domestic armed

group, a member of the governing elite, or some other set of domestic actors. Coups always involve

a departure from the formal or informally accepted rules of transferring power within a particular

country. The use of force may be overt, such as fighting in the capital, or may come in the form of

tacit support by the military and security apparatus of the power grab.25

A change in power within a military junta, if it involves the threat or use of force, is also considered

a coup. Social revolutions and popular uprisings are generally not considered that, unless at some

point a group of actors threatens or uses force to remove the government in place. Unsupported

assassinations, where the perpetrator or perpetrators lack the basic organization or resources to take

power, are not considered a coup. Likewise, where the forcible ouster of a regime is accomplished

solely by foreign actors, the case is not considered a coup.

Our data on coup detats is original. It is constructed from a database on political leaders

Archigos (Goemans et al., 2004). The database codes the identity of all leaders in 164 countries in

the world. It includes information about the manner in which leaders assumed and left office. We

look at how power is transferred between two leaders to identify the set of events that may qualify

as coups. We first code a variable to identify all irregular exits by a leader, then we code a number

of additional variables to distinguish between the different types of irregular exits in the record. An

irregular exit occurs when constitutional or customary provisions for power changeover in a country

are not observed. Additional variables tell us whether the case involved the use of force, whether

force was used or merely threatened, whether the military, rebels, government insiders, or foreignerswere behind the events.

Based on this information, we identify all instances of coup detat that match the definition given

above. In terms of raw numbers, the dataset produces 233 distinct coup events between 1960 and

2004.

There exist other definitions of the coup detat.26 There are also a number of coup datasets.27 We

note briefly what we share in common and how we differ from existing definitions and data collection

strategies.

The basic insight that we share with the literature is the the emphasis on the use or threat of force

in effecting regime change, and the notion that the transfer of power should violate consitutional or

customary procedure. We disagree with some approaches in some or all of the following ways.

First, we do not include unsuccessful coups. Coup plots and attempts are difficult to establish

systematically and independently of potentially questionable claims and interpretations by govern-

25Where the use of power is less than obvious, we need specific evidence that a threat was actually made to concludethat a coup has taken place.

26One alternative definition of the coup detat, due to McGowan and Johnson (2003), sees coups as ... events in whichexisting regimes are suddenly and illegally displaced by the action of relatively small groups, in which members of themilitary, police, or security forces of the state play a key role, either on their own or in conjunction with a number of civilservants or politicians.

27Classical treatments such as Jackman (1978) aside, a recent paper by Belkin and Schofer (2003) features a datasetwith 339 coup events, attempted or successful, between 1945 and 2000. Alesina et al. (1996) present data on 112 countries,between 1960 and 1982. Their source of coup data is the Jodice and Taylor 1983 World Handbook of Social and PoliticalIndicators. Londregan and Poole (1990) use the same source. McGowan (2003) has data on 48 countries in Sub-SaharanAfrica from 1946 to 2001. The data includes successful coups, unsuccessful coups, and plots to overthrow a leader. Perhapsthe most recent effort is the coup events data by Monty G. Marshall and Donna Ramsey at the Center for Systemic Peace.

9


10/20

ments. We do allow, however, leaders to be in office for a brief period of time, such as a week. One

could decide to call a brief tenure in office by a coup leader a failed coup. Indeed, sometimes existing

datasets seem to effectively do that. We feel that keeping power even briefly is a coup. Because we

record the time a coup leader (as well as any leader) keeps office, our data allows for other judgement

calls.

Second, we do not require the transfer of power to result in substantially new policies to qualify

as a coup event. We feel that whether or not the new regime adopts new policies is a dependent

variable in its own right. Our dataset is sufficiently general to allow other scholars to define the

policy changes they may want to study and to ask which coups result in those changes and which

ones do not.

Third, we allow the perpetrators of the coup detat to be members of the government security

apparatus but also to be members of the government itself (Daud Khan to Taraki in Afghanistan),

or rebel forces battling the government (Habre to Deby in Chad). The reason is that we feel the

emphasis on extra-constitutionality and use of force associated with coups traditionally should cover

such events. Some existing datasets claim to adopt a more narrow view of the coup detat, stating

that only takeovers by government insiders (as opposed to insurgents) qualify. However, we havefound those claims to be inconsistently applied. For example, in about one third of the cases in which

we find that rebels effected the government takeover, the Marshall and Marshall dataset concurs that

a coup detat has taken place. The authors never clarify why these cases count but not the other

two thirds of similar transfers of power we uncover.28 Thus, it is not clear why the power transitions

from Gamsakhurdia to Ioseliani in Georgia 1992 and from Nabiyev to Iskandrov in Tajikistan 1992

are not cases of a coup detat. We aim to be always consistent in applying our basic definition to all

cases in the record. That said, we do code specific variables to distinguish the different seizures of

executive power so we know when the rebels vs. the military acted, for example. We are able to test

our key claims against alternative definitions, finding, for example, that it does not make a difference

for our purposes whether rebel-caused power transfers are counted as coups. Researchers would be

able to make these types of judgement calls themselves with our data.

There are other disagreements in the literature.29 We feel that our approach adds some clarity

and flexibility to difficult definitional choices surrounding the coup. We also b elieve that, to the

extent that we systematically go over all leadership changes in the large cross-section of countries we

study, we capture cases that some datasets may have overlooked. The degree of correlation between

our data and other sources ranges between 65 % and 82 %, suggesting that, while we contribute to

the literature on the topic, we still share substantial amount of agreement with previous work.

The Post-Coup Election

Data on elections comes from an original dataset as well. The dataset codes all national-level elections

in 141 countries (Presidential, Legislative, and Parliamentary), together with a variety of attributes.

Some of the attributes coded allow us to determine whether an election was competitive. A com-

petitive election is defined as one in which: (1) political opposition is allowed; with more than one

candidate allowed to run for office; (2) multiple parties are allowed; (3) the office of the incumbent

leader is contested. The first two conditions are perhaps commonsensical, the third requires some

clarification. While many countries hold elections and aspire to be or appear democracies, some

elections absolve the office of the de facto leader of the country from political competition. Jordans

28See the Center for Systemic Peace.29Another definitional conundrum is whether the speed which power is seized should matter. Luttwak (1969), for example,

emphasizes the idea of speed and surgical precision as being central to a coup, and something that distinguishes the coupdetat from an insurgency. Again, we believe that it is best to provide an inclusive set of candidate events while providing

a separate variable to capture the speed of the power transition.

10


11/20

decision to start holding legislative elections in 1989 was hailed with a degree of enthusiasm by pro-

ponents of democratization but no election since then has questioned the ultimate authority of the

king. We do not consider those to be competitive elections, in the sense that we require competition

to involve or concern the ultimate source of power and authority in the country.

We put together the coup and elections data to generate our unit of analysis - the time to election

after a successful coup. We construct coup-spells as the basic building block of our data. A country

enters a coup spell in the year it experiences a successful coup. The country exits the coup spell

when it holds a competitive election. This may happen in the year of the coup, some years later,

or we may still be waiting to see an election. The latter case, in the language of duration analysis,

is the case of right-hand censoring. A country which is currently in a coup spell and experiences a

fresh coup has its current spell censored and enters a fresh coup-spell.

Our dependent variable is the failure of the coup (coup-fail), and we allow this process to have

an underlying duration. Conceiving of the problem in this way allows us to deal appropriately with

cases that are censored.

The two datasets we put together give us information on 139 countries, observed between 1960

and 2004, and yield 202 distinct coup-spells. The average duration of a spell is 8 years, and the rangeis between under a year, to 35 years and more.30

From Coups to Elections: Before and After the Cold War

We want to know how changing international attitudes have changed the time to elections after

a successful seizure of executive power. Before we look at some of the trends, we deal with one

potential objection to our approach. Do the elections we consider really have a bite? We want to

think of the holding of elections as a significant step toward democratization. If it turned out that

post-coup elections did not amount to much, by, for example, allowing coup-leaders to effectively

install themselves in power, our findings may not mean what we believe they do.

We compile a set of characteristics of post-coup elections to speak to this point. The elections weconsider represent a significant step toward political liberalization and not merely a smokescreen to

cover the power ambitions of strongmen.

First, we use the elections data we have to ask whether or not the incumbent leader of the country

in question was replaced as a result of the vote. Table 1 illustrates what we find. We tabulate averages

of the quantities we want to compare for all post-coup elections held in our dataset. We also break

down the comparison by pre- and post-Cold War to check for possible trends in the data. We find

that nearly half (49 %) of all contests (55 of them held after 1991) end with the incumbent leader

being replaced by a newcomer. The pre-1991 statistic is close to two-thirds (66 %). While post-1991

elections seem to be somewhat less prone to fire the executive from office, they still represent a

significant rate of leadership turnover. If the incumbent loses their job half of the time, the contest

can hardly be described as a make-believe sham designed to keep the coup-leaders in charge.31

Second, we ask whether the actor who lead the original coup attempt is out of power after

the elections we consider. We find that in 65 % to 79 % of cases, this is true. These numbers,

again reinforce our sense that these elections can be characterized as a meaningful step away from

dictatorship.

Third, we exploit another variable we have information on to learn about the competitiveness

of the p ost-coup election. We look at the fraudulent-or-free verdicts international election observers

pass on contests they monitor. Since electoral observation took off only after the end of the Cold

30In 2004, Syria and Lybia had two of the longest-running coup spells in our data.31We average rate of leadership turnover for all 1,881 elections in our data (whether preceded by a coup or not) is 1/3,

which means that post-coup elections exceed the average in this respect.

11


12/20

War, we really have little information on this variable prior to 1990. What we do have, though,

is that in the 42 out of 55 total post-coup elections monitored since 1991, international observers

agreed the elections were substiantially free 80 % of the time. This statistic reinforces the notion of

post-coup elections as a genuine liberalizing step.

Finally, we check the liberalization trends recorded by well-known datasets of democratization to

see how post-coup elections fare in that respect. We find that the Polity IV data reports between 2

and 4 point positive change in their Polity2 measure of political liberties.

Table 1: The Bite of Elections After Coups: Different Measures of Liberalization Over Time

Incumbent Replaced Coup-Leader Out Fraud-Free Polity2 Change

Post-CW 49 % 65% 79% 2.04n=55 n=49 n=42 n=55

CW 66 % 79% N/A 3.74n=47 n=47 n=47

The numbers reported by Table 1 show some negative overtime trends in in post-coup elections.

Still, the table indicates that elections after the Cold War result in non-trivial movement in the

direction of political liberty and representative government. Furthermore, our focus on competitive

elections after the coup is not meant to suggest that such contests necessarily amount to a transition

to democracy (which seems to be true about half of the time) but that adopting elections is one,

very important step in an incremental process of democratization.32

We next construct a simple graphical illustration of the overtime variation in our dependent

variable. We look at all successful instances of coup detat in our data, and we look at the five-year

period following the seizure of executive power to see how often competitive elections are observedwithin the chosen time-period. Figure ?? shows the result, by historical period. While the holding

of post-coup elections was the exception before 1991 (observed in 20 % of the cases), after the Cold

War it is the rule (elections occur in 68 % of the cases). We next turn to identifying what explains

the greater incidence of elections after 1991.

4 Analysis

Our identification strategy is based on the observation that the end of the Cold War produced an

exogenous shock in the Wests willingness to tie various benefits to progress toward democracy.

While the Soviet Union was around, there were powerful reasons to look the other way when countryleaders suspended constitutional liberties and rigged elections. After the Soviet Union disappeared,

for reasons arguably unrelated to the propensity of countries to experience free elections, the decision-

making calculus in the powerful power centers in the West changed. The United States and the

European Union moved swiftly to adopt formal and informal rules that called for free elections as

one condition for good relations with the West. Development aid is one important benefit countries

around the world value when it comes to the relations with the west. Aid has the advantage of being

easily measurable for a large set of countries. There are other benefits other countries care about

such as military assistance, alliances, investment. We do not use measures of these in this paper but

we expect the findings to be similar.

32We use the Cheibub and Gandhi (2004) procedural definition of democracy and accompanying dataset to derive this

number.

12


13/20

It is important to note that we do not claim that democratic conditionality, the act of tying ben-

efits to democtatic benefits to be consistemtly applied by different donors against different countries.

Rather, we claim that there is a marked difference between the overall degree to which the West as a

whole is willing to tie aid to democracy before and after the demise of the Soviet Union. Our research

design exploits the shift in this average propensity to tie benefits to democracy before and after the

end of the Cold War. We also confine the analysis to the years prior to 2001 to avoid picking up the

effect of the terrorist attacks against the United States (and any changes to coniditionality that may

have brought about), as well as to avoid picking up the effect of a growing Chinese presence in the

developing world.

We divide our sample into two sets of (country-coup-spell) years: those observed before 1991 and

those observed after. Some spells occur solely before or after 1991. Some spells, 31 out of the total

of 202 we have, span both periods and are present in both of our samples. Our choice of regression

model for the grouped-duration data we have is ordinary logit (Beck, Katz and Tucker, 1998).

We estimate a model in which the dependent variable is whether or not a country experienced

competitive elections in year. Only countries that experienced a coup, and so are in a coup-spell,

participate. Our main independent variable is dependence on Western aid, defined as the ratio of thetotal aid receipts reported to the OECD and the countrys GDP in year. What we expect to find is

that aid dependence should have no effect on whether elections are held before the end of the Cold

War but should be positively related to the occurrence of elections after that.

We include a number of covariates in the estimation. First of all, because we want to know

whether the process of adopting elections is path-dependent, we include a measure of the number of

years since the coup. It maybe that the longer a country remains in a coup-spell, the less likely it is

to adopt elections as leaders consolidate their power base, or become more unpopular and therefore

more wary of a competitive contest. It could also be that the effect goes the other way, with the need

to rebuild legitimacy through competitive elections going up over time. Second, we include a measure

of whether or not the country was an electoral democracy when the coup occurred. The idea is to

see whether institutions are sticky: if countries that have a tradition of electing their government

revert to having elections faster, that would be evidence for the residual bite of institutions.

Next, we include a logge measure of GDP per capita as a way of controlling for the level of

modernization of the country. It maybe that domestic pressures to adopt elections again are higher

in societies at higher level of socio-economic development. Especially because aid dependence and

GDP per capita may be correlated, we need to include this variable as a covariate in the model. We

also include a dummy for whether or not the country is a former French colony. This is an effort to

pick up the presence of French troops. We know that unlike other colonial powers France has been

willing to station troops in colonies and use them in support of against the government in place.

While the dummy captures French military involvement only noisily, it should help distinguish the

effect of the special French policy in troop use abroad.Finally, we code for whether or not the coup began during the Cold War, and include this

variable as a covariate in the p ost-1991 sample. If, as we argue, the expectation to hold elections

was an exogenous shock to the system in 1991, then countries with coup-leaders in power as of 1991

would tend to differ from countries experiencing coups later. Whereas coup-leaders after 1991 could

anticipate the need to quickly move to elections, coup detats conceived before that time did not

have the same foresight. Because post-Cold War leaders are more likely to start a coup in conditions

where the median voters preferred point is an acceptable post-coup outcome, we would expect to

see those leaders move to adopt elections more quickly than the group perpetrating a coup before

that time. We would expect to find that coups that began prior to the shock to the system would

be less likely to result in elections, all else equal.

13


14/20

Table 2: Summary Statistics of Main Variables.

Standard

Mean Deviation Minimum Maximum

Post-Coup Elections 0.068 0.253 0 1

Post-Cold War 0.253 0.435 0 1

Aid Dependence 0.093 0.117 0 1.096

Years Since Coup 7.372 7.650 0 35

Pre-Coup Electoral Democracy 0.209 0.407 0 1

Wealth Per Capita 6.409 1.163 3.898 10.173

French Colony 0.364 0.481 0 1

Coup Starts in Cold War 0.904 0.294 0 1N = 1, 335

Table 3: The Effect of Aid Dependence on the Decision to Hold A Post-Coup Election. Logit,dependent variable post-coup competitive election in year. Standard errors are in parentheses. p < 0.10,p < 0.05,p < 0.01, two-tailed tests. Calculations performed in STATA 10.

pre-CW post-CW (I) post-CW (II)

Aid Dependence 0.267 3.274 3.349

(1.823) (1.263) (1.244)

Years Since Coup -0.082 -0.093(0.036) (0.028)

Pre-Coup Electoral Democracy 0.387 0.526 0.543(0.345) (0.444) (0.443)

Wealth Per Capita 0.251 0.157 0.153(0.141) (0.202) (0.202)

Former French Colony -0.637 0.822 0.832

(0.406) (0.346) (0.346)

Coup Starts in Cold War -0.914 -1.032

(0.571) (0.455)

Constant -4.243 -3.148 -3.159

(1.067) (1.315) (1.316)

N of observations 996 339

n of countries 57 48

Log-Likelihood/df -176.212 -122.448 -122.503

14


15/20

Table 2 presents basic descriptive characteristics of the the data. Table 3 shows results of the logit

analysis. The first column reports the pre-1991 results, and the second and third columns both report

the post-1991 results. The one difference between the second and third columns is the inclusion of

the time-dependence covariates in the second model and their exclusion from the third model. As

expected, aid dependence has no consistent relationship with the speed with which a country moves

to the ballot box after a coup pre-1991. This is in contrast to the aftermath of the Cold War (more

specifically, the 1991-2001 period), when higher aid dependence predicts a faster onset of competitive

elections. This result is consistent with our theoretical expectation.

The years elapsed since the country entered a coup spell have the hypothesized negative effect

on the likelihood of elections, but only for the pre-1991 period. While we find the absence of path-

dependence somewhat surprising after 1991, it could be that the relatively shorter time-frame over

which we observe countries for the latter period provides fewer observations than we need to pick up

a path-dependence that is in fact present in the data.

Dropping the path-dependence control in the post-Cold War sample allows us to more clearly pick

out the effect of whether or not the coup started pre-1991 on how fast elections will be held (years

since coup and the period in which the coup began are significantly correlated). We find a strongnegative association for coups pre-dating 1991 and the likelihood that elections will be adopted

post-1991. This is consistent with our hypothesis that international pressure tends to encourage

different kinds of coups in more recent times. When coup-leaders understand the preferences of the

international community, they are under pressure to undertake coups in conditions where the median

voters preference is more closely aligned with their own preferences. International pressure is both

the reason they select to commit a coup in those circumstances and pressure is the reason they move

to hold elections (rather than rule single-handedly, as they would have in the counterfactual pre-1991

scenario).

Perhaps surprisingly, we find no evidence that whether or not the coup detat is undertaken

against an electoral democracy makes a difference for how quickly electoral institutional are adopted.

It maybe that the occurrence of a coup marks cases where democratic institutions are weak and not

institutionalized, or it maybe that some other explanation is behind this result. We feel that further

research is needed to say more about the stickiness of democratic institutions after a successful coup

detat.

We also find, using level of economic development as a proxy for modernization, that greater

modernization does not predict faster adoption of competitive elections after a coup. The signs of

the coefficients on the lagged and logged GDP is positive in all three models, which is consistent with

the hypothesized effect, but the result is never significant. The non-finding in this case may be due

to attrition on the economic development variable in the set of countries experiencing coups. While

the mean GDP per capita in countries that are in a coup spell is 1,404 dollars (1995 constant terms),

the mean for all countries observed between 1960 and 2001 in the data is 5,235 dollars. Thus, itmay be the case that the variation in the economic development variable that remains in the sample

selected into a coup spell is insufficient to identify the true effect of modernization on the adoption

of elections.

Finally, our findings on how former French colonies differ from the rest in terms of time to elections

after a coup detat is consistent with our expectations. Prior the the end of the Cold War, former

French colonies in a coup-spell were neither more nor less likely to move to elections. This changes

post-1991, with former French colonies being much more likely to adopt competitive elections than

other countries following a coup. We believe that this result may be due in part to strong economic

conditionality coming from Paris (not otherwise picked by the aid dependence variable), but also

possibly, due to an overt or covert threat to use force against successful coup leaders who defy the

15


16/20

directions to call voters to the ballot box.

5 Discussion

Decline in Coups

Our data shows a pronounced over time change in the time it takes a country to hold elections after

a successful coup detat. If our argument is right, and changing international pressure is responsible

for the effect, we would also expect to find a decline in overall coup levels. Per the argument we

develop, pressure to hold elections after a coup makes the prospect of seizing power less attractive

to actors who may otherwise be capable of grabbing power. A successful coup leader after 1991 may

be threatened by international pressure sufficiently to hold elections, or punishment may be too low

to make governing without elections unattractive, but in both cases the payoff they get should be

lower than the one in the counterfactual case of no international repercussions for overthrowing a

government and governing dictatorially. We do not seek to test propositions related to the onset of

the coup detat because we feel that a proper test would exceed the constraints we face. We can still

use our data to plot the over time variation in the incidence of successful coup detats. Figure ??

shows two trends. The bars indicate the number of coups in a given year. The line represents the

number of countries with coup-installed leaders.

The overtime decline in the incidence of coups is evident. The popularity of coups peaked at

the height of the Cold War between 1960 and 1980, with some years recording 10 or more extra-

constitutional seizures of executive power. Before 1991, there was not a single year on record in which

a coup did not succeed at least once. After the end of the Cold War, some years record no coups,

and the maximum number of events we see in a single year does not come close to the maximum

observed in the earlier period.

While the decline in coup-activity is broadly consistent with our argument, we need to make two

points. First, while we still believe that the end of the Cold War is the best proxy for an overall

decline in the tolerance toward coups in the West, we are aware that U.S. policy started to shift

even in the 1980s, and especially so in the Americas. In that sense, some of the decline recorded in

the 1980s is likely consistent with our argument even if it illustartates that the way we measure the

timing and scope of the external shock is somewhat crude.

Second, the spike in coup-activity in 1991 and 1992 is primarily an artefact of the new (unstable)

countries entering the system with the dissolution of the Soviet Union, and of thr overthrow of

dictators previously supported by the Soviets (such as Najibullah in Afghanistan).

Selection Effects

Not all states are equally likely to experience coups. The conditions that give rise to the coup detat

pose a challenge for any attempt to build and test a causal theory of why coups occur and what

happens after one has taken place. This study is no exception to the rule. That said, there are

reasons to be confident that the arguments and the evidence offered tap into a causal rather than

spurious relationship.

The paper seeks to accomplish two goals: (1) demonstrate that coups need not lead to dictator-

ship; (2) that international pressure has caused a more rapid onset of elections after successful coups

in the p ost-Cold War world.

With respect to the first point, selection effects are irrelevant. The mere observation that post-

1991 elections are much more frequent after coups is sufficient to demonstrate the inadequacy of

existing theorizing about what happens after the coup detat.

16


17/20

Selection effects may have a bite with respect to the second point. In particular, we may be

worried that many factors change with the end of the Cold War apart from increased willingness

to tie benefits to democratization by the international community. For example, one may believe

that domestic publics have become disenchanted with dictatorial rule, resulting in domestic pressure

on successful coup-leaders to hold elections and also discouraging the incidence of coups.33 We do

not disagree that such dynamics may be at work. However, in and by itself, the presence of other

changing covariates is not a concern for our estimates unless these covariates are both omitted from

the regression and correletated to the main indendent variable of interest, aid dependence. We would

need to believe, for example, that the level of domestic demand for elected government is related

systematically to levels of aid dependence. We see no particular reason to believe that. While we

cannot rule out all competing explanations, and so endogeneity remains a potential concern, neither

can we readily identify an available explanation that would challenge our identification strategy.

It is also important to note that our argument acknoweledges the self-selection dynamics at work

in the assignment of countries with different local conditions to coups. We argue that, in the presence

of pressure, the types of coups that are least likely to occur are those where international pressure to

hold elections is especially strong, and coups are most likely in cases in which their leaders stand tobenefit from putting the median voter in charge. Our theoretical argument leads us to believe that

we may be able to uncover the causal effects of international pressure on post-coup, the non-random

assignment of countries to coups nothwithstanding. Our model predicts that the type of conditions

that give rise to coups post-1991 will be different from the conditions that caused coups in the earlier

period. We are able to find support for this argument by controlling for the time of the origin of

the coup in the post-Cold War sample. The careful theoretical modelling of selection dynamics we

engage in helps us approach and interpret the evidence better, and the empirical support we find for

the argument gives us confidence that our modelling strategy is appropriate.

6 ConclusionWhen a countrys government is overthrown by a set of actors whose main claim to power is guns

and the ability to deploy violence, there is typically little to cheer about. The coup detat has

been and remains the single most important threat to democracy around the world (the spreading

popularity of autogolpes notwithstanding).34 People often die in such transitions, repressions and

political vendettas are sometimes part of the aftermath, as may be a fall in investment and economic

growth.

Yet, depending on the specific circumstances, a coup can sometimes provide few reasons for

condemnation. When President Yala of Guinea-Bissau was finally overthrown in a bloodless coup

on 14 September 2003, the international community saw little by way of loss. Yalas corrupt, heavy-

handed style of government had few supporters among the countrys donors. When the leaders of thecoup promised to return the country to constitutional rule and elections, international indifference

turned to praise and support. Our results indicate that this event is part of a norm, a systematic

trend in the data, and the exception to the rule. Recent coups tend to lead to elections, and do so

sooner rather than later.

Our results fall well short of an endorsement of the coup detat. But they do point out to two

ways in which these types of transfers of power are less worrisome today than before. First, while the

holding of competitive elections is not democracy, it is a step in the right direction. Todays coup

33Hagopian and Mainwaring (2005) talk about lack of support among the civilians for for military rule as one explanationfor the declining incidence of the coup detat in Latin America.

34We have data on auto-golpes that leads us to believe that autogolpes are nowhere nearly as prevalent as coups, both

before and after the end of the Cold War.

17


18/20

leaders are doing much more to bring their countries closer to democracy than their predecessors in

the 1960s, for example.

Second, because insisting on the holding of elections ex-post would reduce the ex-ante benefit of

seizing power, our findings imply that the incidence of coups should recede. This, given the often

unpalatable consequences of the unconstitutional seizure of executive power, should be a good thing.

Our findings have implications on what the international community can and cannot do to help

democracy around the world. While democracy remains importantly a home-grown institution,

democratic norms have a far better chance of taking root in a country if some minimum procedural

trappings of democratic government can be maintained over time. The worlds poorest countries pose

a difficult terrain for representative political institutions. We argue that the international community

can help in important ways those countries guard their emerging representative institutions from the

dangers of the irregular transfer of power. The actual gain for democracy may be slow and fiful to

materialize in some cases, but this comes with the underlying diffuculty of the task.

18


19/20

References

Acemoglu, Daron and James Robinson. 2006. Economic Origins of Dictatorship and Democracy.Cambridge University.

Alesina, Alberto, Sule Ozler, Nouriel Roubini and Phillip Swagel. 1996. Political Instability andEconomic Growth. Journal of Economic Growth 1:189211.

Beck, Nathaniel, Jonathan N. Katz and Richard Tucker. 1998. Taking Time Seriously: Time-Series-Cross-Section Analysis with a Binary Dependent Variable. American Journal of Political Science42(4):12601288.

Belkin, Aaron and Evan Schofer. 2003. Toward a Structural Understanding of Coup Risk. Journalof Conflict Resolution 47:594620.

Boix, Charles and Susan Stokes. 2003. Endogenous Democratization. World Politics 55:517549.

Cheibub, Jose Antonio. 2006. Presidentialism, Parliamentarism, and Democracy. New York: Cam-bridge.

Cheibub, Jose Antonio and Jennifer Gandhi. 2004. Classifying Political Regimes: A Six FoldClassification of Democracies and Dictatorships. Yale University.

Coufoudakis, Van. 1977. The European Economic Community and the Freezing of the GreekAssociation, 1967-1974. Journal of Common Market Studies 16:114131.

Deutsch, Karl. 1961. Social Mobilization and Political Development. American Political ScienceReview 55:493514.

Dunning, Thad. 2004. Conditioning the Effects of Aid: Cold War Politics, Donor Credibility, andDemocracy in Africa. International Organization 58:409423.

Finer, S. E. 1988. The Man on Horseback: The Role of the Military in Politics. Boulder, CO:Westview.

Goemans, Hein, Kristian Skrede Gleditsch, Giacomo Chiozza and Jinhee L. Choung. 2004. Archigos:A Database on Political Leaders. University of Rochester and University of California San Diego.

Version 1.4.

Hagopian, Frances and Scott Mainwaring. 2005. The Third Wave of Democratization in Latin Amer-ica: Advances and Setbacks. New York: Cambridge University.

Huntington, Samuel. 1968. Political Order in Changing Societies. New Haven, CT: Yale.

Jackman, Robert. 1976. Politicians in Uniform: Military Governments and Social Change in TheThird World. The American Political Science Review 70:10781097.

Jackman, Robert. 1978. The Predictability of Coups dtat: A Model with African Data. TheAmerican Political Science Review 72:12621275.

Johnson, Thomas, Robert Slater and Pat McGowan. 1984. Explaining African Military CoupsdEtat, 1960-1982. The American Political Science Review 78:622640.

LaCouture, Jean. 1970. The Demigods: Charismatic Leadership in the Third World. New York:Alfred Knopf.

Le Vine, Victor. 2004. Politics in Francophone Africa. Boulder, CO: Lynne Rienner.

Lipset, Seymor. 1959. Some Social Requisites of Democracy: Economic Development and PoliticalLegitimacy. American Political Science Review 53:69105.

Londregan, John B. and Keith T. Poole. 1990. Poverty, The Coup Trap, and the Seizure of ExecutivePower. World Politics 42(2):151183.

Luttwak, Edward. 1969. Coup dEtat: A Practical Handbook. New York: Alfred Knopf.

McGowan, Pat and Thomas Johnson. 2003. African Military Coups dEtat and Underdevelopment:

A Quantitative Historical Analysis. The Journal of Modern African Studies 47:594620.

19


20/20

McGowan, Patrick. 2003. African Military Coups dEtat, 1956-2001: Frequency, Trends and Distri-bution. Journal of Modern African Studies 41:339370.

Naftali, Timothy, ed. 2001. John F. Kennedy: The Great Crises. New York: Norton.

Needler, Martin. 1966. Political Development and Military Intervention in Latin America. TheAmerican Political Science Review 60:616626.

OKane, Rosemary. 1981. A Probabilistic Approach to the Causes of Coups dEtat. British Journalof Political Science 11:287308.

Pevehouse, Jon C. 2002. With a Little Help from My Friends: Regional Organizations and theConsolidation of Democracy. American Journal of Political Science 46(3):611626.

Przeworski, A. and F. Limongi. 1997. Modernization - Theories and Facts. World Politics49(2):155183.

Trinkunas, Harold. 2005. Crafting Civilian Control of the Military in Venezuela: A ComparativePerspective. Chapel Hill, NC: University of North Carolina.

Wang, T. Y. 1998. Arms Transfers and Coups dEtat: A Study on Sub-Saharan Africa. Journalof Peace Research pp. 659675.

Yalman, Nur. 1968. Intervention and Extrication: The Officer Corps in the Turkish Crisis. In TheMilitary Intervenes: Case Studies in Political Development, ed. Henry Bienen. New York: RussellSage Foundation.

Zuk, Gary and William Thompson. 1982. The Post-Coup Military Spending Question: A PooledCross-Sectional Time-Series Analysis. American Political Science Review 76:6074.

20

elections after the coup d'etat.pdf

Documents